[12:19] <zul> heylo
[12:59] <BenC> does udev load ide modules in initramfs, or is it something else?
[03:05] <Keybuk> Kamion: I'm going to assume you'd like me to hold off the s-b uploads until Flight 2 is out? :p
[03:05] <Kamion> s-b?
[03:07] <Keybuk> streamlined-boot
[03:08] <Keybuk> largely moving around stuff in rcS to put things back into the right order
[03:08] <Kamion> ah - yes please, hold off
[03:08] <Keybuk> *nods* that was my plan, just wanted to check you weren't "damn it all, put it in!" :p
[03:09] <Keybuk> I'll start working on getting networking to come up automatically now ;)
[03:09] <Nafallo> wfm :-P
[03:09] <Nafallo> but then I have network-manager installed ;-)
[04:22] <BenC> anyone had a chance to build the cd with -8.10 kernels yet?
[04:22] <Kamion> yes, already done
[04:22] <BenC> working well?
[04:22] <Kamion> install CDs only; live CDs need to switch to simplifiedlivecd or something else as yet undetermined, in progress
[04:23] <BenC> ah, the livecd is what I was wondering about, with the cloop change
[04:23] <Kamion> oh, yes, it does get past the previous failure and the cloop appears uncorrupt now
[04:23] <Kamion> on i386, anyhow
[04:24] <BenC> excellent
[04:25] <BenC> any other issues that I need to be aware of?
[04:25] <BenC> before I start pulling in upstream changes for next kernel, I want to make sure I don't need to push out a quick -8.11 :)
[04:25] <Kamion> install CD's doing fine so far, up to the partitioner
[04:25] <Kamion> will let you know
[04:25] <BenC> ok, thanks
[04:26] <Kamion> nothing else I've heard of yet
[04:56] <Kamion> BenC: new live CD spews a bunch of warnings (it's using unionfs too, which seems a bit flaky)
[04:56] <Kamion> but it does manage to boot
[04:58] <BenC> unionfs was slightly broken for us in breezy, but mdz wanted it brought back for dapper
[04:58] <BenC> wait, I might be thinking of another fs
[04:58] <BenC> anyway, glad to hear it boots!
[04:58] <BenC> if you can get a ppc64 livecd made, I'll test it :)
[05:00] <Kamion> on its way, will be a little while yet
[05:20] <crispin> Keybuk, BenC, when I get home should I first try the new kernel, or is there more debugging you want me to do in the old one ?
[05:21] <BenC> crispin: new kernel, I am almost sure it will fix your problem
[05:21] <crispin> BenC: cool, thanks :-)
[05:35] <Mithrandir> yeah, unionfs oops a bit too much, but works-ish.
[05:51] <fabbione> unionfs should die hard
[05:52] <Mithrandir> it's a lot prettier as a solution for the live cd than devmapper is.
[05:52] <jbailey> fabbione: We're going to paint "unionfs" and "xen" on your ceiling above your bed, so it's the first thing you think about in the morning.. ;)
[05:52] <Mithrandir> just fix the bugs.
[05:53] <fabbione> jbailey: hahahah
[05:53] <fabbione> Mithrandir: i know it is... except it will OOPS on ppc
[05:53] <fabbione> and kill your machine
[05:53] <Mithrandir> fabbione: Just Fix The Bugs.
[05:53] <Mithrandir> :-)
[05:53] <fabbione> Mithrandir: no way.. the code is crap
[05:54] <Mithrandir> fabbione: then reengineer the code.
[05:54] <Mithrandir> the idea is sound
[05:56] <fabbione> Mithrandir: if you like it so much you can do it :) i am sure you can and have the 31337 sk1775 for it
[05:56] <Mithrandir> fabbione: WFM. :-)
[05:56] <fabbione> kernel
[07:42] <crispin> Keybuk, BenC, bad news, the new kernel doesn't help
[07:43] <BenC> damn, ok
[07:44] <Keybuk> try booting with break=premount, then do
[07:44] <Keybuk> echo ide-generic > /etc/modprobe.d/test
[07:44] <Keybuk> echo sata_nv >> /etc/modprobe.d/test
[07:44] <Keybuk> scripts/init-premount/udev
[07:44] <Keybuk> mountroot
[07:44] <Keybuk> --
[07:44] <Keybuk> and see whether that works
[07:46] <crispin> Keybuk: should that be with the 'blacklist' at the front like yesterday ?
[07:46] <Keybuk> oh, yes
[07:46] <Keybuk> please :)
[07:47] <crispin> ok, it hung for a while saying "Waiting for root file system"
[07:47] <crispin> and where does this mountroot script live ?
[07:48] <crispin> oh, and I still get this "dev 0 not supported, ignoring" message
[07:49] <Kamion> '. conf/initramfs.conf; . scripts/functions; . scripts/$BOOT' should give you the mountroot function
[07:49] <crispin> Kamion: hi! and thanks
[07:49] <Keybuk> how long a while?
[07:49] <Kamion> hey, how goes
[07:50] <crispin> Keybuk: about 15-20 seconds
[07:50] <Keybuk> k..
[07:50] <crispin> mountroot now says "ALERT! /dev/sda1 does not exist. Dropping to shell"
[07:50] <Keybuk> yeah, thought it would
[07:50] <Keybuk> can you cat /proc/modules
[07:50] <crispin> Kamion: yeah, good, we should meet for a drink sometime
[07:51] <crispin> Keybuk: kde_cd / cdrom / generic / amd74xx / sata_sil / libata / scsi_mod / fbcon / tileblit / font / bitblit / softcursor / capability / commoncap
[07:52] <Keybuk> right
[07:52] <Keybuk> BenC: it definitely looks like sata_sil is bust to me
[07:52] <crispin> err.. ide_cd obviously :-)
[07:53] <BenC> was this working before?
[07:53] <crispin> yep, breezy with same setup is perfect
[07:53] <BenC> sata_sil still?
[07:54] <BenC> is this a PATA drive, or SATA drive?
[07:54] <Kamion> crispin: absolutely
[07:54] <crispin> I believe so, the lsmod and dmesg output are in the bug
[07:55] <crispin> BenC: 160Gb SATA drive hanging off the SATA pci card, and 2 cdrom drives hanging off an old-school IDE interface on the m'board
[07:55] <BenC> let me get the bug number for reference again
[07:56] <crispin> 20807
[07:56] <crispin> Keybuk: oh, that reminds me, sorry about that comment I made on the bug, I hadn't really made my situation clear
[08:10] <Keybuk> BenC: I have another user with exactly the same problem as crispin
[08:10] <Keybuk> sata_sil, same error message, etc.
[08:11] <crispin> I've just copied the dmesg output, and attached it to the bug
[08:11] <crispin> (the output after the scripts/init-premount/udev command)
[08:11] <Keybuk> right
[08:12] <crispin> that was a slow and boring task :-(
[08:12] <Keybuk> heh
[08:13] <crispin> oh, you don't get the numbers before each line, but I'm sure you live without the :-)
[08:13] <crispin> *them
[08:15] <Keybuk> BenC: mehdi has the same bug as crispin
[08:16] <BenC> howdy
[08:18] <BenC> mehdi, crispin: what kernel flavor are you guys using?
[08:19] <BenC> and can you guys build kernels, or would you rather I build something for you?
[08:20] <mehdi> hi
[08:20] <crispin> BenC: the -686, I can build kernels
[08:20] <crispin> argh, no, this box is a -k7 machine
[08:22] <BenC> I'm going to build a sata_sil.ko for you to test
[08:23] <BenC> it reverts some changes that came from libata git, which basically implemented sata_silo local irq handler, as opposed to the libata generic one
[08:24] <mehdi> i can build kernel if you want
[08:24] <mehdi> i already have build my own 2.6.15-6-k7
[08:24] <mehdi> to have a working kernel
[08:24] <mehdi> because since this version the ubuntu kernel don't boot on my box
[08:25] <mehdi> ok
[08:26] <BenC> so -6-k7 worked for you, and -7 and -8 broke?
[08:28] <BenC> MODPOST takes so long :/
[08:33] <BenC> http://people.ubuntu.com/~bcollins/sata_sil.ko
[08:33] <BenC> that's for 2.6.15-8.k7
[08:33] <BenC> -8-k7
[08:34] <BenC> replace /lib/modules/2.6.15-8-k7/kernel/drivers/scsi/sata_sil.ko with that one
[08:34] <BenC> rerun depmod -a
[08:34] <BenC> and reboot, let me know if it works
[08:34] <BenC> oh wait
[08:34] <crispin> how about the initrd ?
[08:34] <BenC> before reboot, run "update-initramfs -u -k 2.6.15-8-k7"
[08:35] <mehdi> ok i just replace th existant one ?
[08:35] <mehdi> ok
[08:37] <crispin> BenC: bingo! it boots!
[08:42] <crispin> now I just need to get my network card started on boot ....
[08:44] <BenC> crispin: sweet!
[08:44] <BenC> echo "auto eth0" >> /etc/network/interfaces
[08:44] <crispin> yeah, just done that :-)
[08:44] <crispin> BenC: thanks very much for such a speedy fix
[08:45] <zul> hah hah..BenC broke it...;)
[08:45] <BenC> no problem
[08:45] <crispin> heh :-)
[08:45] <BenC> zul: nah, jgarzik broke it, but I can't tell him that because I merged code that wasn't meant for wide spread use :)
[08:46] <crispin> now to see whether the clock runs fast, and whether I can unmount my Nokia 770 without errors :-)
[08:47] <mehdi_> i'am back
[08:47] <BenC> medhi: work I assume?
[08:47] <crispin> BenC: at least you now know the code you merged doesn't work in all cases :-)
[08:48] <mehdi_> no can you give me the path where i must copy le modules ?
[08:48] <BenC> you have to install linux-image-2.6.15-8-k7 first
[08:48] <zul> BenC: hehe
[08:49] <BenC> then copy it to /lib/modules/2.6.15-8-k7/kernel/drivers/scsi/sata_sil.ko
[08:49] <BenC> and run "update-initramfs -u -k 2.6.15-8-k7"
[08:54] <mehdi_> it work
[08:55] <BenC> excellent, thank you both for testing
[08:55] <BenC> next kernel will have this code, so you are safe to upgrade
[08:55] <mehdi_> ok thank you very much
[09:10] <mehdi_> i have one question why the acx modules was been removed in the 2.6.15 series ?
[09:10] <mehdi_> it's maybe not the good channel to awser this ?
[09:19] <BenC> they are there
[09:29] <BenC> any ppc64 live cd's yet?
[09:29] <BenC> also, if anyone has info on the unionfs oopses I can look into that aswell
[09:32] <Kamion> BenC: no, need help from lamont on powerpc
[09:32] <BenC> what's up with ppc?
[09:34] <Kamion> initramfs generation complication
[09:34] <Kamion> we need to get the livefs build process to generate two initramfses, not just one like on all the other arches
[09:34] <BenC> 32-bit vs 64-bit?
[09:35] <Kamion> right
[09:35] <BenC> is the livecd build system different from breezy now?
[09:36] <Kamion> yes, we've switched to SimplifiedLiveCD
[09:36] <Kamion> it turned out to be prohibitively difficult to hack the existing casper to support d-i running on initramfs
[09:36] <BenC> is there no way to include both sets of modules in one initramfs?
[09:36] <Kamion> dude, it'll be huge - the powerpc bootloader already has size problems
[09:37] <BenC> ah, true, most of the initramfs is modules
[09:37] <Kamion> we shipped two separate initrds for the breezy powerpc live CD too - it's fine
[09:37] <Kamion> actually easier that way
[09:37] <Kamion> (imo)
[09:37] <BenC> do you include a lot of ide stuff in the initramfs?
[09:38] <BenC> most of the pmac ide drivers are built-in to the kernel
[09:38] <Kamion> that's up to Mithrandir nowadays :)
[09:38] <Kamion> it just uses initramfs-tools now
[09:38] <Kamion> with magic casper hooks
[09:38] <Kamion> so in fact I guess it's up to Adam
[09:39] <BenC> what's the schedule for flight 2 look like now?
[09:40] <Kamion> "before Kamion goes to sleep"
[09:40] <Kamion> (hopefully)
[09:40] <Kamion> when exactly I go to sleep is still up for debate
[09:48] <jbailey> Kamion: What's your record for total-time-awake so far? =)
[11:24] <crispin> BenC: If I am replying to a bug in the NEEDINFO case, how do I get it out of that state ? Is the "Accept bug" option (this is the clock too fast bug)
[11:27] <Nafallo> crispin: the assigne will change the state when he doesn't need more info to fix the bug basically :-)
[11:28] <BenC> I already did it, I though
[11:28] <BenC> changed it to pending upload
[11:28] <BenC> yeah, 20807, pending upload
[11:29] <crispin> BenC: this is 11304 I'll just comment and leave it then
[11:31] <BenC> ah, ok
[11:32] <BenC> yeah there was two bugs, one was fixed (double-speed), the other I cannot reproduce (fast)
[11:33] <BenC> crispin: you have this bug?
[11:33] <crispin> yep
[11:33] <BenC> crispin: if so, can you come in #ubuntu-kernel?
[11:33] <crispin> sure